MODELING OF SPEECH SIGNALS USING AN OPTIMAL NEURAL-NETWORK STRUCTURE-BASED ON THE PMDL PRINCIPLE

Although neural networks have already found many successful applicatio ns in speech signal processing, it is difficult to compare their perfo rmance with that of the traditional methods objectively, In this corre spondence, we address this problem by using the predictive minimum des cription length (PMDL) principle to select the optimal network size in modeling speech signals.